How Temporary Were Canada s Temporary Foreign Workers?

Transcription

1 Catalogue no. 11F0019M No. 402 ISSN ISBN Analytical Studies Branch Research Paper Series How Temporary Were Canada s Temporary Foreign Workers? by Elena Prokopenko and Feng Hou Release date: January 29, 2018

2 How to obtain more information For information about this product or the wide range of services and data available from Statistics Canada, visit our website, You can also contact us by at telephone, from Monday to Friday, 8:30 a.m. to 4:30 p.m., at the following numbers: Statistical Information Service National telecommunications device for the hearing impaired Fax line Depository Services Program Inquiries line Fax line Standards of service to the public Statistics Canada is committed to serving its clients in a prompt, reliable and courteous manner. To this end, Statistics Canada has developed standards of service that its employees observe. To obtain a copy of these service standards, please contact Statistics Canada toll-free at The service standards are also published on under Contact us > Standards of service to the public. Note of appreciation Canada owes the success of its statistical system to a long standing partnership between Statistics Canada, the citizens of Canada, its businesses, governments and other institutions. Accurate and timely statistical information could not be produced without their continued co operation and goodwill. Published by authority of the Minister responsible for Statistics Canada Minister of Industry, 2018 All rights reserved. Use of this publication is governed by the Statistics Canada Open Licence Agreement. An HTML version is also available. Cette publication est aussi disponible en français.

3 How Temporary Were Canada s Temporary Foreign Workers? by Elena Prokopenko and Feng Hou Social Analysis and Modelling Division Statistics Canada 11F0019M No. 402 ISSN ISBN January 2018 Analytical Studies Branch Research Paper Series The Analytical Studies Branch Research Paper Series provides for the circulation of research conducted by Analytical Studies Branch staff and collaborators. The Series is intended to stimulate discussion on a variety of topics, such as labour, immigration, education and skills, income mobility, well-being, aging, firm dynamics, productivity, economic transitions, and economic geography. Readers of the Series are encouraged to contact the authors with their comments and suggestions. All the papers in the Analytical Studies Branch Research Paper Series go through institutional and peer review to ensure that they conform to Statistics Canada s mandate as a governmental statistical agency and adhere to generally accepted standards of good professional practice. Acknowledgements The authors would like to thank Ying Gai, Luin Goldring, Delphine Nakache, and analysts at Immigration, Refugees and Citizenship Canada and Employment and Social Development Canada for advice and comments on an earlier version of this paper. Any errors are the responsibility of the authors.

4 Table of contents Abstract... 5 Executive summary Introduction Theoretical, empirical and policy background Theoretical explanations of return migration Empirical evidence Canada s temporary labour migration program The current study Data, measures and methods Data Measures Methods Results The characteristics of Canada s temporary foreign workers during the 1990s and 2000s Canadian residence status over time Multivariate results Conclusion and discussion Appendix References Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

5 Abstract Temporary foreign worker programs have become an increasingly important component of international migration to Western developed countries. However, there is little knowledge on how long foreign workers stay in the host country and what determinants are associated with their migratory trajectories. Using a national longitudinal administrative dataset of temporary foreign workers (TFWs) in Canada, this study examines their length and type of stay in Canada. It further examines the likelihood of staying given individual demographic characteristics, source-country attributes, host-country institutional factors and local community conditions. The results show that the majority of TFWs stayed in Canada only for a short period, while the majority of those who stayed for a long period obtained permanent resident status. Host-country institutional constraints play a dominant role in determining the length and type of stay of TFWs in Canada. Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

6 Executive summary Temporary foreign workers (TFWs) have been playing a growing role in Canada s labour force and immigration system. The length and type of stay of TFWs in Canada have strong implications for the country s immigration and labour policies. This study assesses the distribution of temporary workers among possible post-arrival residential trajectories to determine which TFWs are more likely to return to their source country and which are more likely to stay in Canada. Specifically, this study examines the impact of individual characteristics, source-country attributes, hostcountry institutional factors and local community conditions on the probabilities of TFWs leaving, transitioning to permanent residence or continuing their temporary stay in Canada. Data for this study are from the Temporary Residents File, which contains demographic and permit-related information on all temporary admitted to Canada since In this study, TFWs are defined as individuals aged 18 to 64 at arrival who received a work permit between 1990 and 2009 and whose first admission to Canada was primarily for work purposes. The analysis excludes TFWs who arrived after 2009 to ensure that the TFWs included in the study were observed for at least five years before the most recent year of observation (2014). Results suggest that the majority of TFWs in Canada were temporary, meaning that most left within the first two years after arrival. However, the tendency to stay longer has increased among more recent arrivals. The share of remaining TFWs declined the most in the first two years after the first work permit. By the fifth year, the decline started to level off, when 13% of the 1995-to cohort and 37% of the 2005-to-2009 cohort still remained in Canada. By the 10th year, the share of TFWs remaining in Canada stabilized at about 11% for the 1995-to-1999 cohort and 18% for the 2000-to-2004 cohort. The overwhelming majority of those who stayed over the long term obtained permanent resident status. Government regulations played the leading role in affecting the length and type of stay of TFWs. There were very large differences in the rate of stay in Canada by arrival cohort. These cohort differences were consistent with Canada s increased reliance on TFWs and the expanded pathways for TFWs to gain permanent residence. Work permit categories were also major indicators of the length and type of stay, even after other predictors were controlled for in a multivariate analysis. This suggests that the terms and conditions attached to most work permit types function mostly independently of individual characteristics, source-country attributes and local economic and social conditions. TFWs from countries with lower levels of economic development and social stability were more likely to stay longer in Canada as temporary or to become permanent. However, many of the differences by source country were accounted for by work permit types. Countries with low levels of economic development and social stability were the main sources of TFWs in the Live-in Caregiver Program, Seasonal Agricultural Workers Program, and Low-Skill Pilot. TFWs in all of these programs had a high tendency to stay longer or come back after leaving for a few months. Individual characteristics (including age and sex), regional unemployment rates and local ethnic concentration had a relatively weak association with the length and type of stay in Canada when work permit type and source-country attributes were considered. Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

7 1 Introduction Global labour movement has intensified, diversified and become more fluid. Many Western developed countries have established institutional arrangements that improve access to a global labour pool. Although these arrangements are meant to be temporary, many temporary foreign workers (TFWs) go on to become long-term in the host countries (Martin 2001). Many TFWs intend to engage in a two-step migration process, through which they first come to the host country on temporary work permits, then find ways to obtain permanent resident status (Khoo, Hugo and MacDonald 2008; Nakache and Dixon-Perera 2015; Nakache and Kinoshita 2010). Some host countries, particularly Australia, Canada and the United States, have established pathways for TFWs to transition to legal permanent residence (Gregory 2014; Hou and Bonikowska 2016; Lowell and Avato 2014). While some TFWs transition to permanent resident status, others prolong their stay by extending their work permits or moving between authorized temporary categories (Goldring and Landolt 2012). The rate and length of stay of TFWs can strongly influence the host country s immigration and labour policies. While annual stock and flow statistics of TFWs are published by Immigration, Refugees and Citizenship Canada (IRCC), little is known about the number of TFWs who have extended their stay as temporary or permanent, and how many have left. Understanding the extent and characteristics of return migration among temporary migrants is essential for evidence-based policy making, given the rapid growth of temporary migration to Canada (Government of Canada 2015; CIC 2014). To address these considerations, this study assesses the distribution of TFWs among possible post-arrival residential trajectories, determining which TFWs are more likely to stay in Canada. Specifically, this study examines the impact of individual characteristics, source-country attributes, host-country institutional factors and local community conditions to determine the probabilities of TFWs transitioning to permanent residence or continuing their temporary stay in Canada. This paper is organized into five sections. Section 2 reviews the literature on theoretical explanations and empirical studies of return migration, and on institutional arrangements of TFWs in Canada. Section 3 contains a discussion on the data sources, measures and analytical approaches. Section 4 presents descriptive statistics and results of multivariate analyses. Section 5 presents the conclusions of the paper. 2 Theoretical, empirical and policy background 2.1 Theoretical explanations of return migration A growing body of literature has expanded the scope of migration research, looking beyond the first migration event to subsequent movements return migration to the individual s country of origin, onward migration to a third country, and a circular migration pattern of frequent moves between two countries. Return migration has several theoretical underpinnings (for an overview, see Budnik 2011; Cassarino 2004; Massey and Espinosa 1997). Early economic theories focus on the decision to return as a means to maximize financial returns (Sjaastad 1962; Stark and Bloom 1985). These theories suggest that a potential migrant s decision to leave the host country is based on a balance of costs and benefits (Borjas and Bratsberg 1996; Dustmann 2003). More recent structural theories consider contextual dimensions of return migration, such as institutional arrangements, and value and power structures in the source country that affect how return migrants use the skills and capital they gain abroad (Cassarino 2004; Callea 1986). Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

8 In addition, political stability, conflict and wars also challenge the ability of return migrants to invest savings from abroad (Borjas and Bratsberg 1996). How much capital they gained abroad matters to migrants as much as whether and how they would benefit from it back home. The transnational theory posits that migrants maintain a simultaneous sense of belonging to both the sending and host societies through so-called transnational practices (Waldinger 2015). While some sending states develop policies that engage the diaspora and compel their return (Faist 2010), many migrants keep abreast of developing opportunities and labour market conditions that are favourable for return to their country (Budnik 2011; Cassarino 2004). Diasporic communities can also facilitate the social and legal integration of new entrants (Castles 2002), curbing return. While many theories on permanent examine reasons for return, other theories explore why temporary migrants who are supposedly bound to return may stay instead. Ruhs (2006) summarized how temporary labour programs can promote permanent settlement of temporary migrants: employer dependence on TFWs can become entrenched in the economy; recruitment and social networks that are established over time can perpetuate and facilitate permanent migration; the judiciary and civil society of liberal democracies can pressure states to extend rights and freedoms to migrants; and workers can exercise their own agency to choose temporary or permanent residence, changing their intentions after experiencing the host country. Balaz, Williams and Kollar (2004) proposed that temporary workers may intend for temporary migration to be the first step in a permanent migration from the outset. They also posit that even migrants who intend to eventually leave may want to use permanent resident status to expand employment and mobility opportunities. The role of host-country institutional factors is unquestionably unique to temporary migrants. After arrival, the decision of permanent to leave largely depends on their own intentions. By contrast, the duration of stay among TFWs is jointly determined by both the worker s intentions and the host country s policies and regulations. Specifically, countries regulate the presence of temporary migrants with visa restrictions, qualifying some migrants more than others to extend their stay or apply for permanent residence under the conditions attached to their work permits. Host-country institutional arrangements dictate the legal terms on which temporary migrants can remain in the country, and can compel return migration regardless of individual intentions. 2.2 Empirical evidence Common patterns in subsequent international mobility (return, onward and circular migration) of the foreign-born population in Western countries were identified from a large body of empirical studies. First, these studies show that the probability of return is highest initially after arrival. This probability declines rapidly at first, then eventually smooths out (Aydemir and Robinson 2008; Bijwaard, Schluter and Wahba 2014; Bratsberg, Raaum and Sørlie 2007; Dustmann and Görlach 2015; Nekby 2006). Second, studies show that outmigration is non-random along several individual characteristics (e.g., Constant and Massey 2002; Borjas and Bratsberg 1996; Aydemir and Robinson 2008; Nekby 2006). For instance, Constant and Massey (2002) found that German language fluency reduced the probability of return migration in Germany. Third, the rate of return among migrants also depends on the socioeconomic and political situations of the source country (Borjas and Bratsberg 1996; Bratsberg, Raaum and Sørlie 2007; Massey and Espinosa 1997). To date, most studies of return migration have focused on foreign-born citizens, particularly permanent, of a country. To the knowledge of the authors, no large-scale quantitative analyses have been conducted of the return migration of TFWs. Temporary workers are difficult to identify in the large bodies of administrative data that do not contain the timing of outmigration and the associated covariates. Most studies on temporary rely on small-scale surveys (e.g., Massey and Espinosa 1997). Several small-scale and qualitative studies that focus on subsequent intentions of temporary migrants find strong support for a two-step migration Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

9 process, wherein temporary migration is seen as a precursor to permanent settlement (Khoo, Hugo and MacDonald 2008; Nakache and Dixon-Perera 2015). In a survey of high-skilled temporary workers in Australia, Khoo, Hugo and MacDonald (2008) showed that more than half of the workers who came to Australia intended to eventually settle permanently, and more than half of those who had no such intentions still applied or intended to apply for permanent residence after arriving. Similarly, Nakache and Dixon-Perera (2015) found from their interviews and focus group discussions with 99 participants that although not all migrants arrived in Canada with the intention to stay permanently, many workers changed their minds once in the country. 2.3 Canada s temporary labour migration program In Canada, government policies and regulations structure the options for temporary remaining in Canada as either temporary or permanent. There are separate streams of temporary work permits that each targets specific industries, skill levels or countries involved in international agreements with Canada. Transitioning to permanent residence is possible through several immigration programs, each with its own criteria of acceptance. These institutional factors play an important role in the rate and duration of stay among TFWs. Since 2014, various streams of temporary work permits have been organized into two overarching programs: the Temporary Foreign Worker Program (TFWP) and the International Mobility Program (IMP). The goal of the TFWP is to fill short-term labour shortages. It is a last resort for employers to fill jobs for which qualified Canadians are not available (Government of Canada 2015, p. 1). It includes two of Canada s most long-standing programs the Seasonal Agricultural Workers Program (SAWP) and the Live-in Caregiver Program (LCP) which allow Canadian employers to hire agricultural and domestic workers on a temporary basis. 1 The Stream for Lower-Skilled Occupations, formerly and for the purposes of this report referred to as the Low-Skill Pilot (LSP), was introduced in 2002 and expanded substantially in the late 2000s. Other TFWP participants are mostly high-skilled workers. Employers who hire workers through the TFWP must submit a labour market impact assessment (LMIA), administered by Employment and Social Development Canada (ESDC). The LMIA is designed to ensure that the employer has made all attempts to hire a Canadian worker for the position, in addition to fulfilling other labour market protecting criteria (Elgersma 2014). Most permits are job-, location- and employer-specific, which limits workers mobility once they are in Canada (Government of Canada 2015). The IMP, the second overarching temporary worker program, aims to advance Canada s broad economic and cultural national interest (Government of Canada 2015, p. 1) and is composed of several different programs. 2 The international agreements program targets nationals of countries with which Canada has signed agreements, such as the North American Free Trade Agreement (NAFTA). The reciprocal employment program includes youth work-exchange programs and facilitates exchanges of employees (e.g., athletes and coaches) or organizations that demonstrate reciprocity by employing Canadian workers abroad (IRCC 2014). The significant benefit program helps entrepreneurs, self-employed individuals and intra-company transferees enter the country. The research- and studies-related work permit holders include postdoctoral fellows, international students completing co-op terms, foreign medical who receive work permits with the 1. The size and importance of the LCP have been significantly reduced since changes to the program were made near the end of 2014 under ministerial instructions. 2. This study uses the subcategories used by IRCC until After 2012, the IMP was reorganized into three main subcategories: agreements, Canadian interests, and other IMP work permit holders. Under Canadian interests, there are four subgroups: significant benefit, reciprocal employment, competitiveness and public policy, and charitable or religious work. Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

10 educational institution as their employer, and former international students in Canada who hold postgraduate work permits (IRCC 2015). Other IMP programs include permits for spouses of IMP program participants, performing artists visiting Canada, and emergency repair and service specialists. Unlike the TFWP, IMP permits do not require a LMIA and thus are not directly tied to current labour market conditions. In most cases, IMP workers and their spouses can receive open work permits, allowing them to easily change employers (Government of Canada 2015). TFWs may extend their stay by either becoming permanent or extending their temporary status, but the options for each stream can be very different. At two extremes are the SAWP and LCP. While all live-in caregivers were allowed to apply for permanent residence after two years of full-time work as domestic workers, 3 SAWP workers have no dedicated stream for transition and may be employed only for a maximum period of eight months per year, after which they must leave Canada (Government of Canada 2016; Hennebry 2014). However, SAWP workers may return to Canada each year on a new LMIA and work permit (IRCC 2016). In general, high-skilled TFWs in both the TFWP and the IMP have more pathways to become permanent because Canada s immigrant selection system rewards candidates for human capital assets such as education, Canadian work experience and official language abilities. For low-skilled workers, the primary pathway to permanent residence is through the provincial or territorial nominee program. Each province or territory with such a program 4 creates criteria and programs for admission; unlike the federal streams, some provincial or territorial nominee programs aid transition for low-skilled workers if there is demand in the local labour market (see detailed discussions by Nakache and Blanchard 2014; Nakache and D Aoust 2012). 2.4 The current study Drawing on theoretical explanations and previous studies on return migration of immigrants, this study considers four sets of determinants for how long TFWs stay and whether they extend their stay by transitioning to permanent residence or remain temporary. The determinants are as follows: individual demographic characteristics, source-country socioeconomic environment, host-country institutional factors, and local destination factors. With regard to individual demographic characteristics, younger TFWs may be more likely to seek longer residence in Canada. This is because younger TFWs are less attached to their sending country and can transition to permanent residence more easily through the Federal Skilled Worker program, which penalizes applicants over the age of 35. Female TFWs are more likely than male TFWs to obtain permanent residence and stay longer for two reasons. First, the LCP, which has a dedicated stream for transition to permanent residence, is largely composed of women. Second, many female TFWs arrive as spouses of high-skilled male TFWs who can easily qualify in the permanent residence application. Low-skilled TFWs are much less likely than high-skilled TFWs to bring their spouses. Even if they do, the spouses of low-skilled TFWs are not eligible for open work permits (Nakache and Kinoshita 2010). Source-country economic and social conditions are key determinants of whether TFWs are motivated and have opportunities to stay longer in Canada. In particular, TFWs from countries with low levels of economic development or low levels of social stability may have a greater quality of life in Canada, and thus are likely more motivated to find ways to stay. TFWs from countries that are far away from Canada are more likely to stay longer because of higher travel costs. 3. Since 2014, new caregivers entering the Canadian market are not admitted through the LCP permanent residence stream. They are admitted through the regular process and do not have a guaranteed pathway to permanent residence. They must now apply under two categories those caring for children, and those caring for people with high medical needs. Both categories have an annual cap. 4. To date, all provinces and territories except Quebec (which already administers its immigration programs separately) and Nunavut have a provincial or territorial nominee program. These programs were adopted at different times, starting with Manitoba in 1996, and ending with Ontario in 2007 (IRCC 2012). Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

11 Furthermore, TFWs from countries where English or French is an official language may find it easier to integrate into Canadian society, and thus may be more likely to stay longer than other immigrants. These TFWs would also be more likely to meet the language criteria for permanent resident applications. Host-country institutional factors are also expected to play a major role in how long TFWs stay in Canada and the type of stay. These factors are represented in this study by arrival periods and work permit types. Because Canada is increasingly using TFWs to meet local or occupational labour demand and establishing more pathways of transition to permanent residence, TFWs who arrived during more recent periods should have higher chances of staying longer as temporary or gaining permanent residence (Foster 2012; Hou and Bonikowska 2016). Work permit types matter in three ways. First, specific transition pathways have been designated for certain types of TFWs, such as the LCP (in this study period) and high-skilled TFWs. Second, certain industrial sectors (e.g., farming, retail trade, accommodation and food services, and transportation and warehousing) have become reliant on the permanent inflow and presence of low-skilled TFWs. Since pathways to permanent residence for these workers are limited (Foster 2012; Nakache and Dixon-Perera 2015), these low-skilled workers are more likely to stay longer as temporary than other TFWs. Third, the duration of some categories of TFWs (e.g., the international agreement category and reciprocal employment category) is bound by the arrangements between Canada and the sending countries. These TFWs encounter more restrictions on staying longer in Canada, either as temporary or permanent, than other TFWs. In addition to national institutional factors, local and regional socioeconomic conditions could affect how long TFWs stay in Canada by acting as push or pull factors. A co-ethnic network in the local communities where TFWs work would help TFWs integrate socioeconomically, which might in turn increase their opportunities and desires to stay in Canada longer. Since government services are generally not extended to temporary, these informal communities are particularly important sources of support and assistance for TFWs, especially when navigating the application process for permanent residence. 5 In regions with a tighter labour market, there would be a stronger demand for TFWs. Employers in these regions are therefore more likely to extend TFW employment or sponsor TFW applications for permanent residence. This study examines how strongly these four sets of determinants affect the duration of stay of TFWs in Canada. With multivariate models, this study can account for possible overlapping effects among these predictors and evaluate their relative importance. 3 Data, measures and methods 3.1 Data Data for this study are from the Temporary Residents File (TRF), which is an administrative dataset from IRCC. The TRF contains demographic and permit-related information of all temporary admitted to Canada between 1980 and 2014, including over 1.3 million TFWs. Canadian census data and national statistics from international organizations were also used to derive variables representing attributes of the source country and local conditions of the host country. 5. Nakache and Dixon-Perera (2015) provided an interesting example in the Brandon community. Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

12 This study defines TFWs as individuals aged 18 to 64 at arrival who received a work permit between 1990 and 2009 and whose first admission to Canada was primarily for work purposes. 6 TFWs who arrived after 2009 have been excluded to ensure that included TFWs have been observed for at least five years before 2014, the most recent year of observation. This study counts only unique persons who came to Canada for temporary work purposes. A person with multiple entries is counted only once in this study. It is important to note that the number of TFWs reported in this study is different from counts of TFWs present in Canada in a given year. 3.2 Measures TFW residence status (temporary resident, permanent resident, or absence) is determined annually. Temporary resident status is defined as holding an authorized temporary resident permit, which could be a work permit or another type of temporary permit. Absence is defined as the lack of legal residence status in Canada for more than one consecutive year. This could include TFWs who left Canada for more than one consecutive year or who stayed in Canada as undocumented persons. Because Canada does not collect information on people leaving the country, it is not possible to estimate how many temporary remain in Canada after their permits have expired. In Canada, there are no efforts to systematically estimate the number of foreign without authorized residence status because undocumented immigration has largely been perceived as a non-issue (Goldring, Berinstein and Bernhard 2009). 7 Although some TFWs might return to Canada after an absence of more than one year, the cut-off is imposed to simplify the analysis. Of all TFWs in the study population, roughly 6% were absent for more than one year, after which they returned, typically as temporary. Individual demographic characteristics include age at arrival and sex. Age at arrival is grouped into four categories: 18 to 24, 25 to 34, 35 to 44, and 45 to 64. The socioeconomic environment of the source country (defined as the country of citizenship) is represented as the standard of living, measured by gross domestic product (GDP) per capita, social stability, distance to Canada and official language (English or French, or other). Each country s GDP per capita (adjusted for purchasing power parity) is averaged over 10 years (1990 to 1999 and 2000 to 2009), and three levels of GDP per capita (low, medium and high) are derived for each decade. 8 Social stability is an index from the Worldwide Governance Indicators (Kaufman, Kraay and Mastruzzi 2010). It measures perceptions of the likelihood of social instability and politically motivated violence. The decade average associated with each country is assigned to each TFW by source country and decade of arrival. Values are then grouped into three levels. 9 The distance between the source country and Canada is included as a proxy for costs of transportation. It is expressed in the hours a flight takes between the capital city of Canada and the capital city of the source country, then grouped into three levels. 10 Finally, the source-country official language variable is a binary indicator of whether English or French is an official language in the TFW s source country This excludes TFWs who initially came to Canada as visitors, international students or refugee claimants and received a work permit at the same time or later. 7. Estimates of undocumented foreign range widely from tens of thousands to half a million. See overviews by Ellis (2015); Goldring, Berinstein and Bernhard (2009); and Magalhaes, Carrasco and Gastaldo (2010). 8. All yearly GDP per capita data are based on purchasing power parity, measured in constant 2010 U.S. dollars. Source-country GDP per capita was classified into low, medium and high levels as less than $10,000, $10,000 to $25,000, and more than $25,000 for the 1990s data; and less than $15,000, $15,000 to $30,000, and more than $30,000 for the 2000s data. The data were downloaded from the World Bank (2017). 9. The values of the original scale range from -2.5 (weak) to 2.5 (strong). In this analysis, they are grouped into low (less than 0), medium (0 to less than 0.75), and high (0.75 or more). 10. The levels are defined as short (less than 6 hours), medium (6 to less than 12 hours) and long (12 hours or more). The data were downloaded from the Happy Zebra Travel Tools (2017). 11. The data were taken from the World Factbook (Central Intelligence Agency n.d.). Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

13 Institutional constraints from the host country are represented by foreign worker streams and arrival cohorts. TFWs initially arrived through one of nine streams identified in the study; 12 under the Temporary Foreign Worker Program (TFWP): (1) Live-in Caregiver Program (LCP), (2) Seasonal Agricultural Workers Program (SAWP), (3) Low-Skill Pilot (LSP), and (4) other TFWP; or under the International Mobility Program (IMP): (5) international agreements, (6) significant benefit, (7) reciprocal employment, (8) research- and study-related, and (9) other IMP. These programs are described in Subsection 2.3. Cohort of arrival, which is based on the first year temporary received a work permit, is used to capture broad policy changes toward TFWs. Four cohorts of arrival are defined, indicating TFWs who first arrived between 1990 and 1994, 1995 and 1999, 2000 and 2004, and 2005 and Local destination characteristics include a measure of co-ethnic concentration and regional unemployment rates. Co-ethnic concentration reflects possible social networks and diaspora presence. It is calculated as the ratio of the proportion of from the same source region living in a particular census subdivision (primarily municipality) to the proportion of from the same source region living in Canada. 13 The data for this measure are derived from the 1991, 1996, 2001 and 2006 censuses. The annual unemployment rate is also included to capture the labour market conditions of the province or major metropolitan area where a TFW resided in a given year. Note that unemployment rates at the broad regional level may not fully capture local or occupational labour shortages. Such a detailed analysis is beyond the scope of this study. Values of both local destination variables are categorized as low, medium or high Methods First, the sample characteristics and the length and type of residential stay in Canada among TFWs are examined. The rates of stay for TFWs as permanent and temporary in Canada, by year since arrival, are graphed separately for each arrival cohort. Next, discrete-time multinomial survival models are run to assess the effects of the selected covariates on an individual s residential status. Because of the large sample size, all covariates showed very small p values, making it difficult to interpret the effects of covariates. To simplify interpretation, estimates are made for the proportions of individuals who stayed as temporary, stayed as permanent or emigrated by the end of the 3rd, 5th and 10th years after obtaining the initial work permit for each covariate. This is based on the model estimates while holding the values of other covariates at their means. These estimated proportions are compared with the observed proportions to show the extent to which other included variables accounted for the observed effect of a given covariate. Furthermore, the unique and common contributions of each set of predictors to the overall model pseudo R-squared value are computed as a way to show their relative importance in predicting the length and type of stay of TFWs in Canada (Nathans, Oswald and Nimon 2012). 12. Some TFWs may switch to different streams during their entire stay in Canada. 13. This is based on 14 broad source regions: North America, Central America (mostly Mexico), the Caribbean, South America, Northern Europe, Western Europe, Southern Europe, Eastern Europe, Africa, South Asia, Southeast Asia, East Asia, West Asia, and Oceania and others. 14. The levels of co-ethnic concentration are low (less than 1, i.e., a group s share in the local areas is below the share of the group in Canada as a whole), medium (1 to less than 2.5) and high (2.5 or more, i.e., a group s share in the local areas is 2.5 times higher than its share in Canada as a whole). The levels of regional unemployment rates are low (less than 6%), medium (6% to less than 9%) and high (9% or more). Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

14 4 Results 4.1 The characteristics of Canada s temporary foreign workers during the 1990s and 2000s In Table 1, TFWs are distributed according to the previously described sets of predictors. Most TFWs are younger than 45 years at arrival, and there were more men than women. Most TFWs came from countries with high levels of GDP per capita; however, the share of TFWs from countries with low GDP per capita increased considerably in the late 2000s. From the 1990s to the 2000s, there was also a large change in the distribution of source-country social stability. This was partially because more TFWs came from countries with low GDP per capita. However, it was also the result of a decreased level of social stability in some major source countries because of terrorism threats, as in the United States. Over two-thirds of TFWs are citizens of countries where English or French is an official language, although this share decreased from the late 1990s to the late 2000s. Consistent with these changes, TFWs increasingly came from countries far away from Canada. In terms of work permit type, most TFWs were admitted without requiring a labour-market opinion (i.e., grouped under the broad IMP category by 2014 definitions) during the 1990s and 2000s. During the 2000s, the share of TFWs requiring a labour-market opinion (i.e. grouped under the broad TFWP category by 2014 definitions) increased significantly mainly because of the expansion of the LSP. Within the IMP, the reciprocal employment category (mostly youth exchange) was the largest component, particularly in the late 2000s, followed by the international agreement category. Among the TFWP streams, the category other TFWP (mostly high-skilled workers) surpassed the low-skilled categories, until the late 2000s, when it was overtaken because of the large expansion of the LSP. The local and regional environment surrounding TFWs has also changed. While over 50% of TFWs settled in municipalities with a low level of co-ethnic concentration in the 1990s, only 38% did so in the late 2000s. Furthermore, the share of TFWs who worked in regions with low unemployment rates (less than 6%) rose substantially from 12% in the early 1990s to 67% in the late 2000s. This large increase was attributable to two factors. First, national macroeconomic conditions improved over this period, and the number of provinces with low levels of unemployment decreased. 15 Second, the large increase of TFWs during the 2000s was concentrated in western Canada where unemployment levels were low. 15. For instance, in the early 1990s, only Ontario had an unemployment rate lower than 6% in 1990 among its prime-age workers. In contrast, in the late 2000s, Ontario and the four western provinces had unemployment rates lower than 6% in most years among their prime-age workers. Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

15 Table 1 Distributions of temporary foreign workers by the selected predictors of the length and type of stay in Canada Period of initial arrival 1990 to to to to 2009 percent Age at arrival 18 to 24 years to 34 years to 44 years to 64 years Sex Male Female Gross domestic product per capita in source country Low Medium High Social stability in source country Low Medium High Flight distance Short Medium Long Official language of source country Not English or French English or French Work permit category Temporary Foreign Worker Program Live-in Caregiver Program Seasonal Agricultural Workers Program Low-skill Pilot Other, Temporary Foreign Worker Program International Mobility Program International agreements Significant benefit Reciprocal Employment Research- and studies-related Other, International Mobility Program Local co-ethnic concentration Low Medium High Regional unemployment rate Low Medium High number Total numbers 223, , , ,540 Note: Percentages may not add up to 100.0% because of rounding. Sources: Statistics Canada, Temporary Residents File and Immigrant Landing File, 1990 to Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

16 4.2 Canadian residence status over time To illustrate the detailed trajectories of TFW stays in Canada, Chart 1 plots the share of TFWs who remain in Canada by years since the first work permit and arrival cohort. It shows a common pattern between the four arrival cohorts. In the first two years, the share of remaining TFWs declined the most, then it continued to decline in the following years. By the fifth year, this share started to level off, when 13% (1995-to-1999 cohort) to 37% (2005-to-2009 cohort) of TFWs still remained in Canada. By the 10th year, the share remaining in Canada stabilized at about 11% for the 1995-to-1999 cohort and 18% for the 2000-to-2004 cohort. These overall trajectories consisted of two different components, as shown in Charts 2 and 3. Chart 2 displays the share of TFWs remaining in Canada as temporary, while Chart 3 displays the share of TFWs who obtained permanent residence. The share of TFWs remaining in Canada as temporary declined the most in the first two years, then continued to decline gradually until about the 10th year, eventually nearing zero (Chart 2). By contrast, the share of TFWs who obtained permanent residence increased the most between the second and fifth year after the first work permit. This share then increased slightly until the 10th year (Chart 3). Therefore, the overall trend observed in Chart 1 was most influenced by the large decline in the share of TFWs who stayed as temporary in the first two years. The decline after the second year in Chart 1 is a result of both the continuing decline in the share of temporary and the large increase in TFWs transitioning to permanent before the fifth year. From the 5th year to the 10th year, there was minimal year-to-year change in the proportion of temporary and permanent. After the 10th year, very few TFWs stayed as temporary, so the level of transition to permanent residence determined the eventual share of TFWs remaining in Canada. Table 2 presents the shares of TFWs remaining in Canada as temporary and permanent in the 3rd, 5th and 10th year by the selected predictors. Note that the sample used to compute the rates for the 3rd year includes TFWs who obtained their first work permit at least 3 years ago, the sample for the 5th year includes TFWs who obtained their first work permit at least 5 years ago, while the sample for the 10th year results includes TFWs who obtained their first work permit at least 10 years ago. Since the samples are different, caution should be exercised in comparing the share of TFWs remaining in Canada by the 3rd, 5th and 10th years. Similar to the trajectories observed by arrival cohort, for all other included predictors, the share of TFWs who stayed as temporary remained sizable by the end of the fifth year. After the fifth year, the share of TFWs who became permanent surpassed the share of TFWs who remained temporary, in most cases. By the 10th year, the remaining TFWs overwhelmingly comprised permanent. The observed proportions of TFWs by residential status at the end of the 3rd, 5th, and 10th year after arrival generally met expectations for all the included predictors (Table 2). TFWs who arrived at the prime working age (ages 25 to 44) had a higher rate of staying as temporary or permanent than their younger or older counterparts. Women were more likely than men to become permanent, but the difference between men and women who remained as temporary was small by the fifth year. Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

17 Chart 1 Percentage of temporary foreign workers remaining in Canada, by years since first work permit percent to-1994 cohort 2000-to-2004 cohort Years since first work permit 1995-to-1999 cohort 2005-to-2009 cohort Sources: Statistics Canada, Temporary Residents File and Immigrant Landing File, 1990 to Chart 2 Percentage of temporary foreign workers remaining in Canada as temporary, by years since first work permit percent Years since first work permit 1990-to-1994 cohort 1995-to-1999 cohort 2000-to-2004 cohort 2005-to-2009 cohort Sources: Statistics Canada, Temporary Residents File and Immigrant Landing File, 1990 to Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

18 Chart 3 Percentage of temporary foreign workers remaining in Canada as permanent, by years since first work permit percent Years since first work permit 1990-to-1994 cohort 2000-to-2004 cohort 1995-to-1999 cohort 2005-to-2009 cohort Sources: Statistics Canada, Temporary Residents File and Immigrant Landing File, 1990 to The rate and type of stay in Canada varied significantly depending on source-country attributes. For example, by the fifth year after the first work permit, 42.8% of TFWs from countries with low GDP per capita acquired permanent residence, while 17.7% remained as temporary. In comparison, among TFWs from countries with high GDP per capita, 7.4% acquired permanent residence and 4.5% remained as temporary. Similarly, 37.9% of TFWs from countries with low social stability became permanent by the fifth year, with another 19.4% remaining in Canada as temporary. In contrast, the corresponding rates were 6.5% and 3.6% among TFWs from countries with high levels of social stability. TFWs from countries that are far away from Canada were much more likely to become permanent than TFWs from other countries. However, distance to Canada was not consistently associated with the rate of remaining as temporary. Whether the source country had English or French as an official language made little difference in the rate of staying in Canada. Host-country institutional factors were significant indicators of the rate of staying. Increases in the shares of TFWs who stayed as temporary or became permanent from the late 1990s cohort to the late 2000s cohort are consistent with Canada s growing reliance on TFWs and increasing pathways for TFWs to gain permanent residence. The relatively high share of TFWs becoming permanent in the 1990-to-1994 cohort resulted from a large-scale, one-time backlog clearance program (Hou and Bonikowska 2016). More strikingly, the work permit category was associated with a very large variation in the type and length of stay in Canada. While the majority of live-in caregivers gained permanent residence by the 5th year, only 2% of seasonal agricultural workers did so even by the 10th year.16 The SAWP had the highest share of TFWs staying as temporary among all permit categories. TFWs in the LSP had relatively high shares of staying as both temporary and permanent. In comparison, TFWs in the other TFWP and research- and study-related categories, both generally high-skilled groups, had moderate shares of TFWs staying as permanent and low shares of staying as temporary. 16. A previous study shows that the main avenue for seasonal agricultural workers to obtain permanent residence is the family class (i.e., being sponsored as family members of other permanent or Canadian citizens) after they left Canada (Lu and Hou 2017). Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402

19 Table 2-1 Observed residence status among temporary foreign workers 3, 5 and 10 years after arrival Temporary 3 years since arrival 5 years since arrival 10 years since arrival Permanent Absent Temporary Permanent Absent Temporary Permanent percent Age at arrival 18 to 24 years to 34 years to 44 years to 64 years Sex Male Female Gross domestic product per capita in source country Low Medium High Social stability in source country Low Medium High Flight distance Short Medium Long Official language of source country Not English or French English or French Arrival cohort 1990 to to to to not applicable Sources: Statistics Canada, Temporary Residents File and Immigrant Landing File, 1990 to Left Analytical Studies Research Paper Series Statistics Canada Catalogue no. 11F0019M, no. 402